A technique of assigning patients to treatment and control groups that is based only on chance distribution. It is used to diminish patient selection bias in clinical trials.